Merge pull request #1155 from JohannesPfeifer/prefilter_tests
Enable erroneously disabled unit teststime-shift
MichelJuillard
commit
d41a64e4cb
|
|
@ -2458,7 +2458,7 @@ in @ref{stoch_simul}, if the @code{periods} option is specified. Note that this
|
|||
in @ref{forecast} as the initial point at which the forecasts are computed
|
||||
|
||||
@item
|
||||
in @ref{conditional_forecast} as the initial point at which the conditional forecasts are computed
|
||||
in @ref{conditional_forecast} for a calibrated model as the initial point at which the conditional forecasts are computed
|
||||
|
||||
@item
|
||||
in @ref{Ramsey} policy, where it also specifies the values of the endogenous states at
|
||||
|
|
@ -4480,7 +4480,9 @@ varobs
|
|||
@descriptionhead
|
||||
|
||||
This block specifies @emph{linear} trends for observed variables as
|
||||
functions of model parameters.
|
||||
functions of model parameters. In case the @code{loglinear}-option is used,
|
||||
this corresponds to a linear trend in the logged observables, i.e. an exponential
|
||||
trend in the level of the observables.
|
||||
|
||||
Each line inside of the block should be of the form:
|
||||
|
||||
|
|
@ -4489,7 +4491,7 @@ Each line inside of the block should be of the form:
|
|||
@end example
|
||||
|
||||
In most cases, variables shouldn't be centered when
|
||||
@code{observation_trends} is used.
|
||||
@code{observation_trends} is used.
|
||||
|
||||
@examplehead
|
||||
|
||||
|
|
@ -6661,7 +6663,7 @@ Use @code{plot_conditional_forecast} to graph the results.
|
|||
|
||||
@item parameter_set = @code{calibration} | @code{prior_mode} | @code{prior_mean} | @code{posterior_mode} | @code{posterior_mean} | @code{posterior_median}
|
||||
Specify the parameter set to use for the forecasting. No default
|
||||
value, mandatory option.
|
||||
value, mandatory option. Note that in case of estimated models, @code{conditional_forecast} does not support the @code{prefilter}-option.
|
||||
|
||||
@item controlled_varexo = (@var{VARIABLE_NAME}@dots{})
|
||||
Specify the exogenous variables to use as control variables. No
|
||||
|
|
@ -6770,6 +6772,8 @@ variables between the first and last specified period. If an intermediate period
|
|||
is not specified, a value of 0 is assumed. That is, if you specify only
|
||||
values for periods 1 and 3, the values for period 2 will be 0. Currently, it is not
|
||||
possible to have uncontrolled intermediate periods.
|
||||
In case of the presence of @code{observation_trends}, the specified controlled path for
|
||||
these variables needs to include the trend component.
|
||||
|
||||
|
||||
@end deffn
|
||||
|
|
|
|||
|
|
@ -98,6 +98,13 @@ Copyright: 1995-2007 Kurt Hornik
|
|||
2008-2009 Dynare Team
|
||||
License: GPL-3+
|
||||
|
||||
Files: matlab/missing/stats/corr.m
|
||||
Copyright: 1993-1996 Kurt Hornik
|
||||
1996-2015 John W. Eaton
|
||||
2013-2015 Julien Bect
|
||||
2016 Dynare Team
|
||||
License: GPL-3+
|
||||
|
||||
Files: matlab/missing/corrcoef/corrcoef.m matlab/missing/corrcoef/sumskipnan.m
|
||||
matlab/missing/corrcoef/flag_implicit_skip_nan.m matlab/missing/corrcoef/tcdf.m
|
||||
Copyright: 2000-2005,2008,2009,2011 by Alois Schloegl <alois.schloegl@gmail.com>
|
||||
|
|
|
|||
|
|
@ -1,4 +1,4 @@
|
|||
function oo_=evaluate_smoother(parameters,var_list)
|
||||
function [oo_, Smoothed_variables_declaration_order_deviation_form]=evaluate_smoother(parameters,var_list)
|
||||
% Evaluate the smoother at parameters.
|
||||
%
|
||||
% INPUTS
|
||||
|
|
@ -11,12 +11,17 @@ function oo_=evaluate_smoother(parameters,var_list)
|
|||
% o oo [structure] results:
|
||||
% - SmoothedVariables
|
||||
% - SmoothedShocks
|
||||
% - SmoothedVariables
|
||||
% - SmoothedVariables
|
||||
% - SmoothedVariables
|
||||
% - SmoothedVariables
|
||||
% - SmoothedVariables
|
||||
% - SmoothedVariables
|
||||
% - FilteredVariablesShockDecomposition
|
||||
% - UpdatedVariables
|
||||
% - FilteredVariables
|
||||
% - SmoothedMeasurementErrors
|
||||
% - FilteredVariablesKStepAhead
|
||||
% - FilteredVariablesKStepAheadVariances
|
||||
% o Smoothed_variables_declaration_order_deviation_form
|
||||
% Smoothed variables from the Kalman smoother in
|
||||
% order of declaration of variables (M_.endo_names)
|
||||
% in deviations from their respective mean, i.e.
|
||||
% without trend and constant part (used for shock_decomposition)
|
||||
%
|
||||
% SPECIAL REQUIREMENTS
|
||||
% None
|
||||
|
|
@ -89,4 +94,8 @@ end
|
|||
|
||||
[atT,innov,measurement_error,updated_variables,ys,trend_coeff,aK,T,R,P,PK,decomp,Trend] = ...
|
||||
DsgeSmoother(parameters,dataset_.nobs,transpose(dataset_.data),dataset_info.missing.aindex,dataset_info.missing.state);
|
||||
[oo_]=write_smoother_results(M_,oo_,options_,bayestopt_,dataset_,dataset_info,atT,innov,measurement_error,updated_variables,ys,trend_coeff,aK,P,PK,decomp,Trend);
|
||||
[oo_]=write_smoother_results(M_,oo_,options_,bayestopt_,dataset_,dataset_info,atT,innov,measurement_error,updated_variables,ys,trend_coeff,aK,P,PK,decomp,Trend);
|
||||
|
||||
if nargout==2
|
||||
Smoothed_variables_declaration_order_deviation_form=atT(oo_.dr.inv_order_var(bayestopt_.smoother_var_list),:);
|
||||
end
|
||||
|
|
@ -68,6 +68,9 @@ else
|
|||
end
|
||||
|
||||
if estimated_model
|
||||
if options_.prefilter
|
||||
error('imcforecast:: Conditional forecasting does not support the prefiltering option')
|
||||
end
|
||||
if ischar(options_cond_fcst.parameter_set)
|
||||
switch options_cond_fcst.parameter_set
|
||||
case 'posterior_mode'
|
||||
|
|
@ -107,18 +110,37 @@ if estimated_model
|
|||
data_index = dataset_info.missing.aindex;
|
||||
gend = dataset_.nobs;
|
||||
missing_value = dataset_info.missing.state;
|
||||
[atT,innov,measurement_error,filtered_state_vector,ys,trend_coeff] = DsgeSmoother(xparam,gend,data,data_index,missing_value);
|
||||
trend = repmat(ys,1,options_cond_fcst.periods+1);
|
||||
for i=1:M_.endo_nbr
|
||||
j = strmatch(deblank(M_.endo_names(i,:)),options_.varobs,'exact');
|
||||
if ~isempty(j)
|
||||
trend(i,:) = trend(i,:)+trend_coeff(j)*(gend+(0:options_cond_fcst.periods));
|
||||
[atT,innov,measurement_error,filtered_state_vector,ys,trend_coeff,aK,T,R,P,PK,decomp,trend_addition] = DsgeSmoother(xparam,gend,data,data_index,missing_value);
|
||||
%get constant part
|
||||
if options_.noconstant
|
||||
constant = zeros(size(ys,1),options_cond_fcst.periods+1);
|
||||
else
|
||||
if options_.loglinear
|
||||
constant = repmat(log(ys),1,options_cond_fcst.periods+1);
|
||||
else
|
||||
constant = repmat(ys,1,options_cond_fcst.periods+1);
|
||||
end
|
||||
end
|
||||
trend = trend(oo_.dr.order_var,:);
|
||||
%get trend part (which also takes care of prefiltering); needs to
|
||||
%include the last period
|
||||
if bayestopt_.with_trend == 1
|
||||
[trend_addition] =compute_trend_coefficients(M_,options_,size(bayestopt_.smoother_mf,1),gend+options_cond_fcst.periods);
|
||||
trend_addition = trend_addition(:,gend:end);
|
||||
else
|
||||
trend_addition=zeros(size(bayestopt_.smoother_mf,1),1+options_cond_fcst.periods);
|
||||
end
|
||||
% add trend to constant
|
||||
for obs_iter=1:length(options_.varobs)
|
||||
j = strmatch(options_.varobs{obs_iter},M_.endo_names,'exact');
|
||||
constant(j,:) = constant(j,:)+trend_addition(obs_iter,:);
|
||||
end
|
||||
trend = constant(oo_.dr.order_var,:);
|
||||
InitState(:,1) = atT(:,end);
|
||||
else
|
||||
graph_title='Calibration';
|
||||
if ~isfield(oo_.dr,'kstate')
|
||||
error('You need to call stoch_simul before conditional_forecast')
|
||||
end
|
||||
end
|
||||
|
||||
if isempty(options_.qz_criterium)
|
||||
|
|
@ -178,7 +200,10 @@ end
|
|||
% number of periods with constrained values
|
||||
cL = size(constrained_paths,2);
|
||||
|
||||
constrained_paths = bsxfun(@minus,constrained_paths,trend(idx,1:cL));
|
||||
%transform constrained periods into deviations from steady state; note that
|
||||
%trend includes last actual data point and therefore we need to start in
|
||||
%period 2
|
||||
constrained_paths = bsxfun(@minus,constrained_paths,trend(idx,2:1+cL));
|
||||
|
||||
FORCS1_shocks = zeros(n1,cL,options_cond_fcst.replic);
|
||||
|
||||
|
|
@ -202,17 +227,15 @@ forecasts.controlled_variables = constrained_vars;
|
|||
forecasts.instruments = options_cond_fcst.controlled_varexo;
|
||||
|
||||
for i = 1:EndoSize
|
||||
eval(['forecasts.cond.Mean.' deblank(M_.endo_names(oo_.dr.order_var(i),:)) ' = mFORCS1(i,:)'';']);
|
||||
forecasts.cond.Mean.(deblank(M_.endo_names(oo_.dr.order_var(i),:)))= mFORCS1(i,:)';
|
||||
tmp = sort(squeeze(FORCS1(i,:,:))');
|
||||
eval(['forecasts.cond.ci.' deblank(M_.endo_names(oo_.dr.order_var(i),:)) ...
|
||||
' = [tmp(t1,:)'' ,tmp(t2,:)'' ]'';']);
|
||||
forecasts.cond.ci.(deblank(M_.endo_names(oo_.dr.order_var(i),:))) = [tmp(t1,:)' ,tmp(t2,:)' ]';
|
||||
end
|
||||
|
||||
for i = 1:n1
|
||||
eval(['forecasts.controlled_exo_variables.Mean.' deblank(options_cond_fcst.controlled_varexo(i,:)) ' = mFORCS1_shocks(i,:)'';']);
|
||||
forecasts.controlled_exo_variables.Mean.(deblank(options_cond_fcst.controlled_varexo(i,:))) = mFORCS1_shocks(i,:)';
|
||||
tmp = sort(squeeze(FORCS1_shocks(i,:,:))');
|
||||
eval(['forecasts.controlled_exo_variables.ci.' deblank(options_cond_fcst.controlled_varexo(i,:)) ...
|
||||
' = [tmp(t1,:)'' ,tmp(t2,:)'' ]'';']);
|
||||
forecasts.controlled_exo_variables.ci.(deblank(options_cond_fcst.controlled_varexo(i,:))) = [tmp(t1,:)' ,tmp(t2,:)' ]';
|
||||
end
|
||||
|
||||
clear FORCS1 mFORCS1_shocks;
|
||||
|
|
@ -231,10 +254,9 @@ end
|
|||
mFORCS2 = mean(FORCS2,3);
|
||||
|
||||
for i = 1:EndoSize
|
||||
eval(['forecasts.uncond.Mean.' deblank(M_.endo_names(oo_.dr.order_var(i),:)) ' = mFORCS2(i,:)'';']);
|
||||
forecasts.uncond.Mean.(deblank(M_.endo_names(oo_.dr.order_var(i),:)))= mFORCS2(i,:)';
|
||||
tmp = sort(squeeze(FORCS2(i,:,:))');
|
||||
eval(['forecasts.uncond.ci.' deblank(M_.endo_names(oo_.dr.order_var(i),:)) ...
|
||||
' = [tmp(t1,:)'' ,tmp(t2,:)'' ]'';']);
|
||||
forecasts.uncond.ci.(deblank(M_.endo_names(oo_.dr.order_var(i),:))) = [tmp(t1,:)' ,tmp(t2,:)' ]';
|
||||
end
|
||||
forecasts.graph.title=graph_title;
|
||||
forecasts.graph.fname=M_.fname;
|
||||
|
|
|
|||
|
|
@ -124,7 +124,7 @@ if ~options_.load_mh_file && ~options_.mh_recover
|
|||
trial = 1;
|
||||
while validate == 0 && trial <= 10
|
||||
candidate = rand_multivariate_normal( transpose(xparam1), d * options_.mh_init_scale, npar);
|
||||
if all(candidate(:) > mh_bounds.lb) && all(candidate(:) < mh_bounds.ub)
|
||||
if all(candidate(:) >= mh_bounds.lb) && all(candidate(:) <= mh_bounds.ub)
|
||||
ix2(j,:) = candidate;
|
||||
ilogpo2(j) = - feval(TargetFun,ix2(j,:)',dataset_,dataset_info,options_,M_,estim_params_,bayestopt_,mh_bounds,oo_);
|
||||
if ~isfinite(ilogpo2(j)) % if returned log-density is
|
||||
|
|
@ -157,6 +157,7 @@ if ~options_.load_mh_file && ~options_.mh_recover
|
|||
end
|
||||
if trial > 10 && ~validate
|
||||
disp(['Estimation::mcmc: I''m unable to find a starting value for block ' int2str(j)])
|
||||
fclose(fidlog);
|
||||
return
|
||||
end
|
||||
end
|
||||
|
|
@ -166,7 +167,7 @@ if ~options_.load_mh_file && ~options_.mh_recover
|
|||
else% Case 2: one chain (we start from the posterior mode)
|
||||
fprintf(fidlog,[' Initial values of the parameters:\n']);
|
||||
candidate = transpose(xparam1(:));%
|
||||
if all(candidate(:) > mh_bounds.lb) && all(candidate(:) < mh_bounds.ub)
|
||||
if all(candidate(:) >= mh_bounds.lb) && all(candidate(:) <= mh_bounds.ub)
|
||||
ix2 = candidate;
|
||||
ilogpo2 = - feval(TargetFun,ix2',dataset_,dataset_info,options_,M_,estim_params_,bayestopt_,mh_bounds,oo_);
|
||||
disp('Estimation::mcmc: Initialization at the posterior mode.')
|
||||
|
|
@ -179,6 +180,7 @@ if ~options_.load_mh_file && ~options_.mh_recover
|
|||
else
|
||||
disp('Estimation::mcmc: Initialization failed...')
|
||||
disp('Estimation::mcmc: The posterior mode lies outside the prior bounds.')
|
||||
fclose(fidlog);
|
||||
return
|
||||
end
|
||||
fprintf(fidlog,' \n');
|
||||
|
|
|
|||
|
|
@ -0,0 +1,189 @@
|
|||
function retval = corr(x, y) % --*-- Unitary tests --*--
|
||||
%@info:
|
||||
%! @deftypefn {Function File} {} corr (@var{x})
|
||||
%! @deftypefnx {Function File} {} corr (@var{x}, @var{y})
|
||||
%! Compute matrix of correlation coefficients.
|
||||
%! @anchor{corr}
|
||||
%!
|
||||
%! If each row of @var{x} and @var{y} is an observation and each column is
|
||||
%! a variable, then the @w{(@var{i}, @var{j})-th} entry of
|
||||
%! @code{corr (@var{x}, @var{y})} is the correlation between the
|
||||
%! @var{i}-th variable in @var{x} and the @var{j}-th variable in @var{y}.
|
||||
%! @tex
|
||||
%! $$
|
||||
%! {\rm corr}(x,y) = {{\rm cov}(x,y) \over {\rm std}(x) {\rm std}(y)}
|
||||
%! $$
|
||||
%! @end tex
|
||||
%! @ifnottex
|
||||
%!
|
||||
%! @example
|
||||
%! corr (x,y) = cov (x,y) / (std (x) * std (y))
|
||||
%! @end example
|
||||
%!
|
||||
%! @end ifnottex
|
||||
%! If called with one argument, compute @code{corr (@var{x}, @var{x})},
|
||||
%! the correlation between the columns of @var{x}.
|
||||
%! @end deftypefn
|
||||
%@eod:
|
||||
%
|
||||
% Notes: - the original Octave code has been rewritten to avoid calling cov, since
|
||||
% there is a long-standing incompatiblity between Matlab's cov and Octave's cov
|
||||
% (see https://savannah.gnu.org/bugs/?40751)
|
||||
% - For compatibility with Matlab, the correlation of a constant
|
||||
% is defined as NaN, not 1
|
||||
%
|
||||
% Adapted for Matlab (R) from GNU Octave 4.0.1
|
||||
% Original files: statistics\base\corr.m, statistics\base\cov.m, and packages\stk-2.3.4\misc\mole\corr\corr.m
|
||||
% Original authors: Kurt Hornik <hornik@wu-wien.ac.at> and Julien Bect <julien.bect@supelec.fr>
|
||||
|
||||
% Copyright (C) 1993-1996 Kurt Hornik
|
||||
% Copyright (C) 1996-2015 John W. Eaton
|
||||
% Copyright (C) 2013-2015 Julien Bect
|
||||
% Copyright (C) 2016 Dynare Team
|
||||
%
|
||||
% This file is part of Dynare.
|
||||
%
|
||||
% Dynare is free software: you can redistribute it and/or modify
|
||||
% it under the terms of the GNU General Public License as published by
|
||||
% the Free Software Foundation, either version 3 of the License, or
|
||||
% (at your option) any later version.
|
||||
%
|
||||
% Dynare is distributed in the hope that it will be useful,
|
||||
% but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||||
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||||
% GNU General Public License for more details.
|
||||
%
|
||||
% You should have received a copy of the GNU General Public License
|
||||
% along with Dynare. If not, see <http://www.gnu.org/licenses/>.
|
||||
|
||||
|
||||
if (nargin < 1 || nargin > 2)
|
||||
error('corr needs to be called with 1 or 2 input arguments')
|
||||
end
|
||||
|
||||
% Input validation
|
||||
if (nargin==1 && ~(isnumeric (x) || islogical (x))) || ...
|
||||
(nargin==2 && ~(isnumeric (x) || islogical (x) || isnumeric (y) || islogical (y)))
|
||||
error ('corr: X and Y must be numeric matrices or vectors');
|
||||
end
|
||||
|
||||
if (nargin==1 && ~ismatrix(x)) || (nargin==2 && (~ismatrix(y) || ~ismatrix(x)))
|
||||
error ('corr: X and Y must be 2-D matrices or vectors');
|
||||
end
|
||||
|
||||
if (nargin == 2)
|
||||
if size(y,1)~=size(x,1)
|
||||
error('corr: X and Y must have the same length')
|
||||
end
|
||||
end
|
||||
|
||||
% Special case, correlation with scalar is NaN in Matlab
|
||||
if isscalar(x)
|
||||
if nargin==1
|
||||
retval = NaN;
|
||||
else
|
||||
retval = NaN(size(y));
|
||||
end
|
||||
return;
|
||||
end
|
||||
|
||||
if nargin==2 && isscalar(y)
|
||||
retval = NaN(size(x'));
|
||||
return;
|
||||
end
|
||||
|
||||
n = size (x, 1);
|
||||
x = x - repmat(mean(x),n,1); %demean
|
||||
sx = std(x,1); %standard deviation
|
||||
sx(sx==0)=NaN; %take care of constant vectors
|
||||
|
||||
if (nargin == 2)
|
||||
y = y - repmat (mean (y), n, 1);
|
||||
sy = std (y, 1);
|
||||
sy (sy == 0) = nan;
|
||||
else
|
||||
y = x;
|
||||
sy = sx;
|
||||
end
|
||||
|
||||
c = x'*y;
|
||||
s = sx'*sy;
|
||||
retval = c./(n * s);
|
||||
|
||||
|
||||
end
|
||||
|
||||
|
||||
%@test:1
|
||||
%$test
|
||||
%$ x = rand (10);
|
||||
%$ cc1 = corr (x);
|
||||
%$ cc2 = corr (x, x);
|
||||
%$ t(1)=dassert(size(cc1),[10, 10]);
|
||||
%$ t(2)=dassert(size (cc2),[10, 10]);
|
||||
%$ t(3)=dassert(cc1, cc2, sqrt (eps));
|
||||
%@eof:1
|
||||
|
||||
%@test:2
|
||||
%$ x = [1:3]';
|
||||
%$ y = [3:-1:1]';
|
||||
%$ t(4)=dassert(corr (x, y), -1, 5*eps);
|
||||
%$ t(5)=dassert(corr (x, flipud (y)), 1, 5*eps);
|
||||
%$ t(6)=dassert(corr ([x, y]), [1 -1; -1 1], 5*eps);
|
||||
%@eof:2
|
||||
|
||||
%@test:3
|
||||
%$t(7)=dassert (corr (5), NaN);
|
||||
%$t(8)=dassert (corr([1 2 3],5),NaN(3,1));
|
||||
%$t(9)=dassert (corr(5,[1 2 3]),NaN(1,3));
|
||||
%@eof:3
|
||||
|
||||
%@test:4
|
||||
%$ Test input validation
|
||||
%$ try
|
||||
%$ corr ()
|
||||
%$ t(10) = false;
|
||||
%$ catch
|
||||
%$ t(10) = true;
|
||||
%$ end
|
||||
%@eof:4
|
||||
|
||||
%@test:5
|
||||
%$ Test input validation
|
||||
%$ try
|
||||
%$ corr (1, 2, 3)
|
||||
%$ t(11) = false;
|
||||
%$ catch
|
||||
%$ t(11) = true;
|
||||
%$ end
|
||||
%@eof:5
|
||||
|
||||
%@test:6
|
||||
%$ Test input validation
|
||||
%$ try
|
||||
%$ corr ([1; 2], ['A', 'B'])
|
||||
%$ t(12) = false;
|
||||
%$ catch
|
||||
%$ t(12) = true;
|
||||
%$ end
|
||||
%@eof:6
|
||||
|
||||
%@test:7
|
||||
%$ Test input validation
|
||||
%$ try
|
||||
%$ corr (ones (2,2,2))
|
||||
%$ t(13) = false;
|
||||
%$ catch
|
||||
%$ t(13) = true;
|
||||
%$ end
|
||||
%@eof:7
|
||||
|
||||
%@test:8
|
||||
%$ Test input validation
|
||||
%$ try
|
||||
%$ corr (ones (2,2), ones (2,2,2))
|
||||
%$ t(14) = false;
|
||||
%$ catch
|
||||
%$ t(14) = true;
|
||||
%$ end
|
||||
%@eof:8
|
||||
|
|
@ -59,13 +59,9 @@ end
|
|||
function build_figure(name,cci1,cci2,mm1,mm2,options_,graphoptions)
|
||||
hh = dyn_figure(options_,'Name',['Conditional forecast (' graphoptions.title ,'): ' name '.']);
|
||||
H = length(mm1);
|
||||
h1 = area(1:H,cci1(2,1:H));
|
||||
set(h1,'BaseValue',min([min(cci1(1,:)),min(cci2(1,:))]))
|
||||
set(h1,'FaceColor',[.9 .9 .9])
|
||||
h1 = area(1:H,cci1(2,1:H),'BaseValue',min([min(cci1(1,:)),min(cci2(1,:))]),'FaceColor',[.9 .9 .9]);
|
||||
hold on
|
||||
h2 = area(1:H,cci1(1,1:H));
|
||||
set(h2,'BaseValue',min([min(cci1(1,:)),min(cci2(1,:))]))
|
||||
set(h2,'FaceColor',[1 1 1])
|
||||
h2 = area(1:H,cci1(1,1:H),'BaseValue',min([min(cci1(1,:)),min(cci2(1,:))]),'FaceColor',[1 1 1]);
|
||||
plot(1:H,mm1,'-k','linewidth',3)
|
||||
plot(1:H,mm2,'--k','linewidth',3)
|
||||
plot(1:H,cci2(1,:),'--k','linewidth',1)
|
||||
|
|
|
|||
|
|
@ -19,7 +19,7 @@ function oo_ = shock_decomposition(M_,oo_,options_,varlist)
|
|||
% SPECIAL REQUIREMENTS
|
||||
% none
|
||||
|
||||
% Copyright (C) 2009-2013 Dynare Team
|
||||
% Copyright (C) 2009-2016 Dynare Team
|
||||
%
|
||||
% This file is part of Dynare.
|
||||
%
|
||||
|
|
@ -62,7 +62,7 @@ if isempty(parameter_set)
|
|||
end
|
||||
end
|
||||
|
||||
oo = evaluate_smoother(parameter_set,varlist);
|
||||
[oo,Smoothed_Variables_deviation_from_mean] = evaluate_smoother(parameter_set,varlist);
|
||||
|
||||
% reduced form
|
||||
dr = oo.dr;
|
||||
|
|
@ -77,16 +77,15 @@ A = dr.ghx;
|
|||
B = dr.ghu;
|
||||
|
||||
% initialization
|
||||
gend = eval(['size(oo.SmoothedShocks.' M_.exo_names(1,:),',1)']);
|
||||
gend = size(oo.SmoothedShocks.(M_.exo_names(1,:)),1);
|
||||
epsilon=NaN(nshocks,gend);
|
||||
for i=1:nshocks
|
||||
epsilon(i,:) = eval(['oo.SmoothedShocks.' M_.exo_names(i,:)]);
|
||||
epsilon(i,:) = oo.SmoothedShocks.(M_.exo_names(i,:));
|
||||
end
|
||||
|
||||
z = zeros(endo_nbr,nshocks+2,gend);
|
||||
for i=1:endo_nbr
|
||||
z(i,end,:) = eval(['oo.SmoothedVariables.' M_.endo_names(i,:)]);
|
||||
end
|
||||
|
||||
z(:,end,:) = Smoothed_Variables_deviation_from_mean;
|
||||
|
||||
maximum_lag = M_.maximum_lag;
|
||||
lead_lag_incidence = M_.lead_lag_incidence;
|
||||
|
|
|
|||
|
|
@ -275,7 +275,7 @@ MODFILES = \
|
|||
optimizers/fs2000_w.mod \
|
||||
differentiate_forward_vars/RBC_differentiate_forward.mod \
|
||||
TeX/fs2000_corr_ME.mod \
|
||||
prior_posterior_function/fs2000_prior_posterior_function.mod
|
||||
prior_posterior_function/fs2000_prior_posterior_function.mod \
|
||||
observation_trends_and_prefiltering/MCMC/Trend_loglin_no_prefilt_first_obs_MC.mod \
|
||||
observation_trends_and_prefiltering/MCMC/Trend_loglin_prefilt_first_obs_MC.mod \
|
||||
observation_trends_and_prefiltering/MCMC/Trend_loglinear_no_prefilter_MC.mod \
|
||||
|
|
@ -299,8 +299,7 @@ MODFILES = \
|
|||
observation_trends_and_prefiltering/calib_smoother/Tr_no_prefilter_loglin_calib_smoother.mod \
|
||||
observation_trends_and_prefiltering/calib_smoother/Tr_no_prefil_f_obs_loglin_cal_smoother.mod \
|
||||
observation_trends_and_prefiltering/calib_smoother/Tr_prefilter_loglin_calib_smoother.mod \
|
||||
observation_trends_and_prefiltering/calib_smoother/Tr_prefil_f_obs_loglin_cal_smoother.mod \
|
||||
reporting/example1.mod
|
||||
observation_trends_and_prefiltering/calib_smoother/Tr_prefil_f_obs_loglin_cal_smoother.mod
|
||||
|
||||
XFAIL_MODFILES = ramst_xfail.mod \
|
||||
estim_param_in_shock_value_xfail.mod \
|
||||
|
|
@ -524,7 +523,8 @@ EXTRA_DIST = \
|
|||
observation_trends_and_prefiltering/Trend_exp_model_calib_no_prefilter_common.inc \
|
||||
observation_trends_and_prefiltering/Trend_model_calib_prefilter_common.inc \
|
||||
observation_trends_and_prefiltering/Trend_model_calib_no_prefilter_common.inc \
|
||||
observation_trends_and_prefiltering/Trend_load_data_common.inc
|
||||
observation_trends_and_prefiltering/Trend_load_data_common.inc \
|
||||
observation_trends_and_prefiltering/Trend_no_prefilter_conditional_forecast.inc \
|
||||
optimal_policy/Ramsey/oo_ramsey_policy_initval.mat \
|
||||
optimizers/optimizer_function_wrapper.m \
|
||||
optimizers/fs2000.common.inc \
|
||||
|
|
|
|||
|
|
@ -107,7 +107,7 @@ varobs gp_obs gy_obs;
|
|||
options_.solve_tolf = 1e-12;
|
||||
|
||||
// Metropolis replications are too few, this is only for testing purpose
|
||||
estimation(order=1,datafile=fsdat_simul,nobs=192,loglinear,mh_replic=10000,mh_nblocks=1,mh_jscale=0.8);
|
||||
estimation(order=1,datafile=fsdat_simul,nobs=192,loglinear,mh_replic=2000,mh_nblocks=1,mh_jscale=0.8);
|
||||
|
||||
conditional_forecast_paths;
|
||||
var gy_obs;
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
mode_compute=4,first_obs=1000,loglinear,smoother,forecast=100,prefilter=0,
|
||||
mcmc_jumping_covariance='MCMC_jump_covar',
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mh_nblocks=1,mh_jscale=0.3) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,20 +1,23 @@
|
|||
@#include "../Trend_exp_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
mode_compute=4,first_obs=1000,loglinear,smoother,forecast=100,prefilter=1,
|
||||
mcmc_jumping_covariance='MCMC_jump_covar_prefilter',
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mh_nblocks=1,mh_jscale=1e-4) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params_prefilter');
|
||||
loaded_par=load('orig_params_prefilter');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
loaded_par_full=load('../orig_params');
|
||||
loaded_par_full=load('orig_params');
|
||||
y_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_y',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_y',loaded_par_full.param_names,'exact'));
|
||||
p_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_p',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_p',loaded_par_full.param_names,'exact'));
|
||||
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
mode_compute=4,first_obs=1,loglinear,diffuse_filter,smoother,forecast=100,prefilter=0,
|
||||
mcmc_jumping_covariance='MCMC_jump_covar',
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mh_nblocks=1,mh_jscale=0.3) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,20 +1,23 @@
|
|||
@#include "../Trend_exp_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=2000,
|
||||
mode_compute=4,first_obs=1,loglinear,smoother,forecast=100,prefilter=1,
|
||||
mcmc_jumping_covariance='MCMC_jump_covar_prefilter',
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mh_nblocks=1,mh_jscale=1e-4) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params_prefilter');
|
||||
loaded_par=load('orig_params_prefilter');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
loaded_par_full=load('../orig_params');
|
||||
loaded_par_full=load('orig_params');
|
||||
y_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_y',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_y',loaded_par_full.param_names,'exact'));
|
||||
p_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_p',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_p',loaded_par_full.param_names,'exact'));
|
||||
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=2000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=2000,
|
||||
mode_compute=4,first_obs=1,smoother,mh_nblocks=1,mh_jscale=0.3,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mcmc_jumping_covariance='MCMC_jump_covar',forecast=100,prefilter=0) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,15 +1,18 @@
|
|||
@#include "../Trend_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',
|
||||
mh_replic=2000,mode_compute=4,first_obs=1000,smoother,forecast=100,prefilter=0,
|
||||
mcmc_jumping_covariance='MCMC_jump_covar',
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mh_nblocks=1,mh_jscale=0.3) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,20 +1,23 @@
|
|||
@#include "../Trend_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=2000,mode_compute=4,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=2000,mode_compute=4,
|
||||
first_obs=1,smoother,prefilter=1,
|
||||
mh_nblocks=1,mh_jscale=1e-4,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mcmc_jumping_covariance='MCMC_jump_covar_prefilter',forecast=100) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params_prefilter');
|
||||
loaded_par=load('orig_params_prefilter');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
loaded_par_full=load('../orig_params');
|
||||
loaded_par_full=load('orig_params');
|
||||
y_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_y',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_y',loaded_par_full.param_names,'exact'));
|
||||
p_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_p',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_p',loaded_par_full.param_names,'exact'));
|
||||
|
||||
|
|
|
|||
|
|
@ -1,20 +1,23 @@
|
|||
@#include "../Trend_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=2000,mode_compute=4,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=2000,mode_compute=4,
|
||||
first_obs=1000,smoother,prefilter=1,
|
||||
mh_nblocks=1,mh_jscale=1e-4,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
mcmc_jumping_covariance='MCMC_jump_covar_prefilter',forecast=100) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params_prefilter');
|
||||
loaded_par=load('orig_params_prefilter');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
loaded_par_full=load('../orig_params');
|
||||
loaded_par_full=load('orig_params');
|
||||
y_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_y',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_y',loaded_par_full.param_names,'exact'));
|
||||
p_forecast_100_periods=loaded_par_full.orig_params(strmatch('const_p',loaded_par_full.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par_full.orig_params(strmatch('g_p',loaded_par_full.param_names,'exact'));
|
||||
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=0,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=0,
|
||||
mode_compute=4,first_obs=1,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
loglinear,smoother,forecast=100,prefilter=0) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
@ -17,3 +20,6 @@ y_forecast_100_periods=loaded_par.orig_params(strmatch('const_y',M_.param_names,
|
|||
p_forecast_100_periods=loaded_par.orig_params(strmatch('const_p',M_.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par.orig_params(strmatch('g_p',M_.param_names,'exact'));
|
||||
|
||||
@#include "../Trend_diagnostics_ML_common.inc"
|
||||
|
||||
@#include "../Trend_no_prefilter_conditional_forecast.inc"
|
||||
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=0,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=0,
|
||||
mode_compute=4,first_obs=1000,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
loglinear,smoother,forecast=100,prefilter=0) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
@ -17,3 +20,5 @@ y_forecast_100_periods=loaded_par.orig_params(strmatch('const_y',M_.param_names,
|
|||
p_forecast_100_periods=loaded_par.orig_params(strmatch('const_p',M_.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par.orig_params(strmatch('g_p',M_.param_names,'exact'));
|
||||
|
||||
@#include "../Trend_diagnostics_ML_common.inc"
|
||||
|
||||
@#include "../Trend_no_prefilter_conditional_forecast.inc"
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
first_obs=1,smoother,loglinear,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
forecast=100,prefilter=1) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
|
|
@ -18,3 +21,4 @@ y_forecast_100_periods=loaded_par.orig_params(strmatch('const_y',loaded_par.para
|
|||
p_forecast_100_periods=loaded_par.orig_params(strmatch('const_p',loaded_par.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par.orig_params(strmatch('g_p',loaded_par.param_names,'exact'));
|
||||
|
||||
@#include "../Trend_diagnostics_ML_common.inc"
|
||||
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../Exp_AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='Exp_AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
first_obs=1000,smoother,loglinear,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
forecast=100,prefilter=1) P_obs Y_obs junk2;
|
||||
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
|
|
|
|||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=0,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=0,
|
||||
mode_compute=4,first_obs=1,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
diffuse_filter,smoother,forecast=100,prefilter=0) P_obs Y_obs junk2;
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
@ -16,3 +19,5 @@ y_forecast_100_periods=loaded_par.orig_params(strmatch('const_y',M_.param_names,
|
|||
p_forecast_100_periods=loaded_par.orig_params(strmatch('const_p',M_.param_names,'exact'))+(options_.first_obs+options_.nobs-1+options_.forecast)*loaded_par.orig_params(strmatch('g_p',M_.param_names,'exact'))
|
||||
|
||||
@#include "../Trend_diagnostics_ML_common.inc"
|
||||
|
||||
@#include "../Trend_no_prefilter_conditional_forecast.inc"
|
||||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_model_no_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=0,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=0,
|
||||
mode_compute=4,first_obs=1000,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
smoother,forecast=100,prefilter=0) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs(M_.params-loaded_par.orig_params)./loaded_par.orig_params)>0.03
|
||||
error('Parameter estimates do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',mh_replic=0,mode_compute=4,
|
||||
first_obs=1,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
smoother,forecast=100,prefilter=1) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
|
|
|
|||
|
|
@ -1,15 +1,18 @@
|
|||
@#include "../Trend_model_prefilter_common.inc"
|
||||
|
||||
estimation(order=1,datafile='../AR1_trend_data_with_constant',
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
estimation(order=1,datafile='AR1_trend_data_with_constant',
|
||||
mh_replic=0,mode_compute=4,
|
||||
filtered_vars, filter_step_ahead = [1,2,4],
|
||||
first_obs=1000,diffuse_filter,smoother,forecast=100,prefilter=1) P_obs Y_obs junk2;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameter estimates do not match')
|
||||
|
|
|
|||
|
|
@ -0,0 +1,39 @@
|
|||
conditional_forecast_paths;
|
||||
var P_obs;
|
||||
periods 1 2 3:5;
|
||||
values 1.01 0.98 1;
|
||||
var Y_obs;
|
||||
periods 1 2 3:5;
|
||||
values 3 3.04 2.98;
|
||||
end;
|
||||
|
||||
conditional_forecast(periods=100,parameter_set=posterior_mode,replic=1000, controlled_varexo=(e_y,e_p));
|
||||
|
||||
plot_conditional_forecast(periods=100) Y_obs P_obs;
|
||||
|
||||
%compare unconditional forecasts
|
||||
cond_forecast=load('conditional_forecasts.mat');
|
||||
if max(abs(cond_forecast.forecasts.uncond.Mean.P_obs(2:end)-oo_.forecast.Mean.P_obs))>1e-8
|
||||
error('Unconditional Forecasts do not match')
|
||||
end
|
||||
|
||||
%compare conditional forecasts; histval here sets initval condition for capital different from steady state
|
||||
initial_condition_states=NaN(size(oo_.dr.ys));
|
||||
initial_condition_states(strmatch('Y_obs',M_.endo_names,'exact')) = oo_.SmoothedVariables.Y_obs(end)-oo_.Smoother.Trend.Y_obs(end);
|
||||
initial_condition_states(strmatch('P_obs',M_.endo_names,'exact')) = oo_.SmoothedVariables.P_obs(end)-oo_.Smoother.Trend.P_obs(end);
|
||||
initial_condition_states(strmatch('junk1',M_.endo_names,'exact')) = oo_.SmoothedVariables.junk1(end);
|
||||
initial_condition_states(strmatch('junk2',M_.endo_names,'exact')) = oo_.SmoothedVariables.junk2(end)-oo_.Smoother.Trend.junk2(end);
|
||||
shock_matrix = zeros(options_cond_fcst_.periods ,M_.exo_nbr); %create shock matrix with found controlled shocks
|
||||
shock_matrix(1:5,strmatch('e_y',M_.exo_names,'exact')) = cond_forecast.forecasts.controlled_exo_variables.Mean.e_y; %set controlled shocks to their values
|
||||
shock_matrix(1:5,strmatch('e_p',M_.exo_names,'exact')) = cond_forecast.forecasts.controlled_exo_variables.Mean.e_p; %set controlled shocks to their values
|
||||
|
||||
y_simult = simult_(initial_condition_states,oo_.dr,shock_matrix,1);
|
||||
|
||||
if max(abs(y_simult(strmatch('Y_obs',M_.endo_names,'exact'),:)'+(options_.first_obs-1+options_.nobs:options_.first_obs-1+options_.nobs+options_.forecast)'*g_y-cond_forecast.forecasts.cond.Mean.Y_obs))>1e-8
|
||||
error('Conditional Forecasts do not match')
|
||||
end
|
||||
|
||||
if max(abs(y_simult(strmatch('P_obs',M_.endo_names,'exact'),:)'+(options_.first_obs-1+options_.nobs:options_.first_obs-1+options_.nobs+options_.forecast)'*g_p-cond_forecast.forecasts.cond.Mean.P_obs))>1e-8
|
||||
error('Conditional Forecasts do not match')
|
||||
end
|
||||
|
||||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_exp_model_calib_no_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../Exp_AR1_trend_data_with_constant',prefilter=0,loglinear,first_obs=1000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='Exp_AR1_trend_data_with_constant',prefilter=0,loglinear,first_obs=1000,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_model_calib_no_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../AR1_trend_data_with_constant',prefilter=0,first_obs=1000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='AR1_trend_data_with_constant',prefilter=0,first_obs=1000,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs;
|
||||
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,12 +1,16 @@
|
|||
@#include "../Trend_model_calib_no_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
calib_smoother(datafile='../AR1_trend_data_with_constant',prefilter=0,
|
||||
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='AR1_trend_data_with_constant',prefilter=0,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs;
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,14 +1,17 @@
|
|||
@#include "../Trend_exp_model_calib_no_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../Exp_AR1_trend_data_with_constant',prefilter=0,loglinear,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='Exp_AR1_trend_data_with_constant',prefilter=0,loglinear,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params)./loaded_par.orig_params))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_exp_model_calib_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../Exp_AR1_trend_data_with_constant',prefilter=1,loglinear,first_obs=1000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='Exp_AR1_trend_data_with_constant',prefilter=1,loglinear,first_obs=1000,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_model_calib_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../AR1_trend_data_with_constant',prefilter=1,first_obs=1000,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='AR1_trend_data_with_constant',prefilter=1,first_obs=1000,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs;
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_model_calib_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../AR1_trend_data_with_constant',prefilter=1,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='AR1_trend_data_with_constant',prefilter=1,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs;
|
||||
load('../AR1_trend_data_with_constant');
|
||||
load('AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -1,13 +1,16 @@
|
|||
@#include "../Trend_exp_model_calib_prefilter_common.inc"
|
||||
options_.filter_decomposition=1;
|
||||
|
||||
calib_smoother(datafile='../Exp_AR1_trend_data_with_constant',prefilter=1,loglinear,
|
||||
addpath('..');
|
||||
generate_trend_stationary_AR1;
|
||||
|
||||
calib_smoother(datafile='Exp_AR1_trend_data_with_constant',prefilter=1,loglinear,
|
||||
// filter_decomposition,
|
||||
filtered_vars, filter_step_ahead = [1,2,4]) P_obs Y_obs junk2;
|
||||
load('../Exp_AR1_trend_data_with_constant');
|
||||
load('Exp_AR1_trend_data_with_constant');
|
||||
@#include "../Trend_load_data_common.inc"
|
||||
|
||||
loaded_par=load('../orig_params');
|
||||
loaded_par=load('orig_params');
|
||||
if max(abs((M_.params-loaded_par.orig_params([1:4,7:8]))./loaded_par.orig_params([1:4,7:8])))>0.03
|
||||
error('Parameters do not match')
|
||||
end
|
||||
|
|
|
|||
|
|
@ -0,0 +1,85 @@
|
|||
function generate_trend_stationary_AR1
|
||||
n_periods=10000;
|
||||
rho_y=0.5;
|
||||
rho_p=0.5;
|
||||
g_y=0.0001;
|
||||
g_p=-0.0001;
|
||||
const_y=2;
|
||||
const_p=2;
|
||||
sigma_y=0.001;
|
||||
sigma_p=0.001;
|
||||
|
||||
orig_params=[rho_y rho_p g_y g_p sigma_y sigma_p]';
|
||||
param_names=char('rho_y','rho_p','g_y','g_p','sigma_y','sigma_p');
|
||||
|
||||
save orig_params_prefilter orig_params param_names
|
||||
|
||||
orig_params=[rho_y rho_p g_y g_p const_y const_p sigma_y sigma_p]';
|
||||
param_names=char('rho_y','rho_p','g_y','g_p','const_y','const_p','sigma_y','sigma_p');
|
||||
|
||||
save orig_params orig_params param_names
|
||||
|
||||
jumping_covariance=diag([1e-8; 1e-8; 1e-16; 1e-16; 1e-8; 1e-8; 1e-12; 1e-12;])^-1;
|
||||
save MCMC_jump_covar jumping_covariance
|
||||
jumping_covariance=diag([1e-8; 1e-8; 1e-16; 1e-16; 1e-12; 1e-12;])^-1;
|
||||
save MCMC_jump_covar_prefilter jumping_covariance
|
||||
|
||||
%% data without constant
|
||||
log_P=zeros(1,n_periods);
|
||||
log_Y=zeros(1,n_periods);
|
||||
junk2_orig=zeros(1,n_periods);
|
||||
for ii=2:n_periods
|
||||
log_P(ii)=rho_p*log_P(ii-1)+sigma_p*randn;
|
||||
log_Y(ii)=rho_y*log_Y(ii-1)+sigma_y*randn;
|
||||
junk2_orig(ii)=0.9*junk2_orig(ii-1)+randn;
|
||||
end
|
||||
%add trend
|
||||
log_P=log_P+g_p*(1:n_periods);
|
||||
log_Y=log_Y+g_y*(1:n_periods);
|
||||
|
||||
Y_obs=exp(log_Y);
|
||||
P_obs=exp(log_P);
|
||||
junk2=exp(junk2_orig);
|
||||
save Exp_AR1_trend_data_no_constant Y_obs P_obs junk2
|
||||
%
|
||||
% [b_p,~,~,~,stats_p] = regress(log(P_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' log(P_obs(1:end-1)')]);
|
||||
% [b_y,~,~,~,stats_y] = regress(log(Y_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' log(Y_obs(1:end-1)')]);
|
||||
|
||||
Y_obs=log_Y;
|
||||
P_obs=log_P;
|
||||
junk2=junk2_orig;
|
||||
save AR1_trend_data_no_constant Y_obs P_obs junk2
|
||||
|
||||
% [b_p,~,~,~,stats_p] = regress((P_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' (P_obs(1:end-1)')]);
|
||||
% [b_y,~,~,~,stats_y] = regress((Y_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' (Y_obs(1:end-1)')]);
|
||||
|
||||
%% data with constant
|
||||
log_P=zeros(1,n_periods);
|
||||
log_Y=zeros(1,n_periods);
|
||||
log_P(1,1)=const_p;
|
||||
log_Y(1,1)=const_y;
|
||||
for ii=2:n_periods
|
||||
log_P(ii)=(1-rho_p)*const_p+rho_p*log_P(ii-1)+sigma_p*randn;
|
||||
log_Y(ii)=(1-rho_y)*const_y+rho_y*log_Y(ii-1)+sigma_y*randn;
|
||||
end
|
||||
%add trend
|
||||
log_P=log_P+g_p*(1:n_periods);
|
||||
log_Y=log_Y+g_y*(1:n_periods);
|
||||
|
||||
Y_obs=exp(log_Y);
|
||||
P_obs=exp(log_P);
|
||||
junk2=exp(junk2_orig);
|
||||
save Exp_AR1_trend_data_with_constant Y_obs P_obs junk2
|
||||
|
||||
% [b,bint,r,rint,stats] = regress(log(P_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' log(P_obs(1:end-1)')]);
|
||||
% [b,bint,r,rint,stats] = regress(log(Y_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' log(Y_obs(1:end-1)')]);
|
||||
|
||||
Y_obs=log_Y;
|
||||
P_obs=log_P;
|
||||
junk2=junk2_orig;
|
||||
save AR1_trend_data_with_constant Y_obs P_obs junk2
|
||||
|
||||
% [b_p,~,~,~,stats_p] = regress((P_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' (P_obs(1:end-1)')]);
|
||||
% [b_y,~,~,~,stats_y] = regress((Y_obs(2:end))',[ones(n_periods-1,1) (2:n_periods)' (Y_obs(1:end-1)')]);
|
||||
|
||||
|
||||
|
|
@ -1,3 +1,8 @@
|
|||
@#include "fs2000.common.inc"
|
||||
|
||||
estimation(mode_compute=2,order=1, datafile='../fs2000/fsdat_simul', nobs=192, mh_replic=0);
|
||||
estimation(mode_compute=2,order=1, datafile='../fs2000/fsdat_simul', nobs=192, mh_replic=0,
|
||||
optim=(
|
||||
'MaxIter',5000,
|
||||
'TolFun',1e-4,
|
||||
'TolX',1e-4)
|
||||
);
|
||||
|
|
|
|||
Loading…
Reference in New Issue